The present invention relates to live attenuated gram-negative vaccine carrier strains which are useful for expression and delivery of heterologous O-antigens (O-PS) from gram-negative pathogens. Said strains are deficient in the expression of homologous O-PS due to a defined genetic modification, preferably a deletion, and, thus, capable of efficiently expressing a desired heterologous O-PS in such a way that it is covalently coupled either to homologous or heterologous LPS core lipid A. The present invention furthermore relates to live vaccine carrier strains containing a heterologous gene or a set of heterologous genes encoding O-PS. Preferably, said strains additionally contain genes necessary for the synthesis of complete smooth heterologous LPS. The present invention also relates to live vaccines comprising said strains, preferably for immunization against gram-negative enteric pathogens.
Gram-negative enteric pathogens are the cause of a variety of diseases presenting with a broad spectrum of symptoms ranging from mild watery diarrhea to severe life-threatening symptoms such as fever, bloody diarrhea, perforation or ulceration of the stomach or intestine, alone or in combination. Examples of such diseases include typhoid fever, shigellosis, cholera, infections with enterotoxinogenic, enteropathogenic, and enterohemorragic Escherichia coli, and infections with Heliobacter pylori and Campylobacter jejuni. 
The first stage of the infectious process occurs at the mucosal surface within the digestive tract. Thus, interfering with this initial stage of infection prior to the onset of symptoms offers a particularly attractive approach. The most effective means by which to accomplish this would be to evoke a local protective immune response through the use of an orally administered vaccine (Mestecky, J. Clin. Immunol. 7 (1987), 265-276; McGhee and Kiyono, Infect. Agents Dis. 2 (1993), 55-73; Walker, Vaccine 12 (1994), 387-400). At present, 2 live oral attenuated vaccines against enteric disease have been licensed for human use these being the Ty21a strain of Salmonella typhi for the prevention of typhoid fever and the CVD103-HgR strain of Vibrio cholerae for the prevention of cholera (Germanier and Furer, J.Infect.Dis, 131 (1975), 553-558; Levine et al., Lancet ii (1998), 467-470).
There exists a large body of evidence indicating that protection against several enteric pathogens, such as S. typhi, E. coli, and Shigella species is associated with the induction of an immune response against cell surface components, specifically the O-antigen moiety of LPS, commonly referred to as O-polysaccharide (O-PS). For example, immunity to shigellosis, subsequent to recovery from either naturally-acquired or experimentally-induced disease is correlated with a substantial rise in serum serotype-specific anti-LPS antibodies (DuPont et al., J.Infect.Dis. 125 (1972), 5-11; DuPont et al., J.Infect.Dis. 12 (1972), 12-16; Herrington et al., Vaccine 8 (1990), 353-357). Furthermore, epidemiological studies have also found that protection against Shigella infections in the field was associated with increased levels of serum anti-LPS antibodies (Cohen et al. , J.Infect.Dis. 157 (1988), 1068-1071). High levels of serum antibodies against Shigella LPS can be detected among individuals residing in areas where such species of Shigella are endemic, presumably acquired by natural exposure and/or infection with these pathogens.
LPS is an essential constituent of the gram-negative outer membrane and may account for up to 70% of the cell surface components. LPS is composed of 3 regions: the innermost being lipid A which is embedded into the phospholipid outer membrane bilayer. The core polysaccharide is attached to the lipid A moiety usually via 2-keto,3-deoxyoctonate (KDO). The core is usually comprised of 5 to 7 sugars. To date, 7 types of core molecules have been identified within the Enterobacteriaceae family and have been named Ra, R1, R2, R3, R4, K-12, and B. Compared with the Enterobacteriaceae, V. cholerae possess an unusual core structure in that it contains fructose and a single KDO molecule in the inner core (Kondo et al., Carbohydrate Res. 231 (1992), 55-64). The biosynthesis of the LPS core is encoded by the rfa locus. Among the Enterobacteriaceae, the rfa and rfb loci appear to be unlinked. In contrast, some evidence exists to suggest a close linkage of at least part of these two loci for V. cholerae (Manning et al., p. 77-94. In Vibrio cholerae and Cholera: molecular to global perspectives (1994). Wachsmuth K., Blake, P. A., and Olsvik Y. (eds.). Washington, D.C.: American Society for Microbiology).
The outermost portion of the LPS molecule is composed of the O-PS which consists of repeating saccharide units of variable length (Luderitz et al., Curr. Top. Membr. Trans. 17 (1982), 79-151; Raetz, Annu. Rev. Biochem. 59 (1990) 129-170). The O-PS region of the LPS molecule confers serospecificity to the bacteria. The LPS molecule interacts closely with other molecules expressed on the outer membrane surface such as porins and other outer membrane proteins (OMP), which determine the permeability of the outer membrane. It is known that the assembly of OMP as well as secretion of proteins from the cell is affected by mutations in the LPS of E. coli (Laird et al., J. Bacteriol. 176 (1994), 2259-2204; Stanley et al., Mol. Microbiol. 10 (1993) 781-787).
Serospecificity is conferred not only by the sugars present in the O-PS but also by their chemical linkage and sequence (Lxc3xcderitz et al., Curr. Top. Membr. Trans. 17 (1982), 79-151). Therefore, the O-PS is highly variable between gram-negative bacterial species whereas the core polysaccharide is relatively constant within a given species or genera (Lxc3xcderitz et al., Curr. Topics in Membranes and Transport 17 (1982), 79-151; Jansson et al., Eur.J.Biochem. 115 (1981), 571-577). For example, the genus Shigella includes a total of 47 known serotypes divided among the 4 predominant pathogenic species which are S. dysenteriae (subgroup A, 12 serotypes), S. flexneri (subgroup B, 13 serotypes) S. boydii (serogroup C, 18 serotypes) and S. sonnei (subgroup D; 1 serotype) (Ewing, In: Ewing W H, ed. Edwards and Ewing""s identification of Enterobacteriaceae fourth edition. New York: Elsevier Sci. Publish. Comp. (1986), 135-172). For example, in S. sonnei, the O-PS consists of a repeated disaccharide unit with 2 unusual sugars, 2-amino-2-deoxy-L-alturonic acid linked to 2-acetamido-4-amino-2,4,6-trideoxy-D-galactose by a 1,4 linkage (Kenne et al., Carbohydrate Res. 78 (1980), 119-126). In contrast, the O-PS of serotype 1 of S. dysenteriae (which is the most common cause of dysentery) is composed of repeating blocks of rhamnose-rhamnose-galactose-N-acetylglucosamine (Ewing and Lindberg, In: Bergan T. (ed) Methods in microbiology vol.14. , Academic Press, London, pp. 113-142). The O-PS of V. cholerae 01 is comprised of 17-18 perosamine subunits each of which is acylated with 3-deoxy-L-glycero-tetronic acid. Quinovosamine has also been found in low concentrations but its location within the O-PS of V. cholerae 01 is unknown (Redmond, FEBS Lett. 50 (1975), 147-149; Kenne et al., Carbohydrate Res. 100 (1982), 341-349).
The enzymes involved in the biosynthesis of enterobacterial O-PS are coded for by the rfb locus. In the case of Shigella species, an additional gene, termed rfc, encodes the O-PS polymerase which functions to polymerize the individual repeat units into chains of varying length. In most Shigella species, the rfb/rfc loci are located on the chromosome (Klena and Schnaitman, Microbiol. Rev. 57 (1993), 655-682). However, in some species of Shigella, all or part of the rfb locus is located on a plasmid episome (Maurelli and Sansonetti, Ann. Rev. Microbiol. 42 (1988), 127-150). An additional gene, termed rfe, which is involved in the synthesis of the enterobacterial common antigen (ECA) is also required for O-PS synthesis in Salmonella species of the O-antigen groups C1 and L (Kuhn et al., FEMS Microbiol. Rev. 54 (1988), 195-222; Mxc3xa4kelxc3xa4 et al., J. Gen. Microbiol. 60 (1970), 91-106), as well as in some serotypes of E. coli and in S. dysenteriae type 1 (Kuhn et al., FEMS Microbiol. Rev. 54 (1988), 195-222; Schmidt et al., J. Bacteriol. 127 (1976), 755-762; Klena and Schnaitman, Microbiol. Rev. 57 (1993), 655-682). A further gene, termed rfp, encodes a galactosyl transferase and is necessary for the production of full-length O-PS in S. dysenteriae type 1 (Klena and Schnaitman, Microbiol. Rev. 57 (1993), 655-682). In addition, serotype conversion can be accomplished via substitution of an O-PS sugar promoted by certain phages lysogenic for Salmonella species and S. flexneri (Clark et al. , Gene 107 (1991), 43-52; Verma et al. Gene 129 (1993) 101).
In the specific case of V. cholerae, the entire rfb locus is chromosomally encoded. Genes involved in perosamine synthesis (rfbABDE), transport of the polymerized O-PS to the cell surface (rfbGHI), and in the transfer of tetronic acid onto the perosamine subunit (rfbKLMNO), are sequentially organized to constitute a single operon. In addition, four genes of unknown function, termed rfbPQRS, constitute the 3xe2x80x2 end of the operon. Directly adjacent to the rfb operon is the rfbT gene which determines the Inaba and Ogawa serospecificity of 01 strains of V. cholerae. It was recently determined that the Inaba serotype strains, are rfbT mutants (Manning et al., p.77-94. In Vibrio cholerae and Cholera: molecular to global perspectives (1994). Wachsmuth, K., Blake, P. A., and Olsvik, xcfx86. (eds.). Washington, D.C.:American Society for Microbiology.
As noted above, the induction of a local intestinal immune response may be the most efficient means by which to prevent infection with a number of enteric pathogens. A proven and effective method by which to accomplish this is through the use of live oral attenuated vaccine strains. Vaccine strains such as S. typhi Ty21a and V. cholerae CVD103-HgR noted above undergo an abortive infectious process thereby inducing an immune response closely resembling that effected by natural infection. The above two strains possess the distinct advantage of being extremely safe in humans (Levine et al., Rev. Infect. Dis. 11 (1989), (Suppl 3), 552-567; Cryz et al., Infect. Immun. 61 (1993), 1149-1151; Levine and Kaper, Vaccine 11 (1993), 207-212).
Safety has been found to be the most difficult attribute to achieve in the development of live oral vaccine strains. Most often, candidate vaccine strains either induce a protective immune response but with an unacceptable rate of adverse reactions or are safe but non-protective (Lindberg, In Vaccine and Immunotherapy. Cryz Jr,S. J. (ed.). New York: Pergamon Press Inc. (1991), pp. 95-112; Levine and Hone, In Vaccine and Immunotherapy. Cryz Jr,S. J. (ed.). New York: Pergamon Press Inc. (1991), pp. 59-72).
Given the above, it is desirable to utilize approved live oral attenuated vaccine strains as carriers for the delivery of heterologous vaccine antigens to the intestinal tract. Attempts to utilize the S. typhi Ty21a strain as a carrier for vaccine antigens has not yielded promising results (Curtiss III, In: New generation vaccines. Woodrow, G. C. and Levine, M. M. (eds.) New York: Marcel Dekker Inc. (1990), pp. 161-188; Cardenas and Clements, Clin. Microbiol. Rev. 5 (1992), 328-342). This in large part can be accounted for by the fact that this strain was developed using a potent chemical mutagen which induced multiple mutations. Therefore, the precise attenuating mutation is unknown. Furthermore, the Ty21a strain replicates poorly in vivo requiring multiple doses of vaccine to be administered. In contrast, the CVD103-HgR vaccine strain was constructed using recombinant DNA technology allowing for the precise genetic lesions to be identified (Ketley et al., FEMS Microbiol. Lett. 111 (1993), 15-22). Furthermore, this strain appears to replicate well in vivo as evidenced by the fact that only a single dose of vaccine is required to induce a high level of immunity against experimental cholera (Levine et al., Lancet ii (1988), 467-470).
Initial attempts to utilize the above strains as carriers envisioned the development of bivalent vaccines. In such a case, the recombinant strain would co-express two O-PS antigens. However, the successful development of such bivalent vaccine strains has proven to be extremely difficult for a variety of reasons, some of which are just becoming apparent. First, experimental data has shown that covalent linkage between the O-PS moiety and LPS core region appears to be a prerequisite for the efficient induction of immunity (Beckmann et al., Nature 201 (1964), 1298-1301; Kuhn et al., FEMS Microbiol. Rev. 54 (1988), 195-222; Attridge et al., Microb. Path. 8 (1990), 177-188; Baron et al., Infect. Immun. 55 (1987), 2797-2801). Second, the co-expression of two O-PS entities often results in the masking of one antigen thereby blunting the immune response (Attridge et al., Microb. Path. 8 (1990), 177-188; Forrest et al., Vaccine 9 (1991), 515-520). Third, the recombinant strain must still fully express the protective antigens associated with the carrier strain. Finally, expression of the foreign antigen should not adversely affect the ability of the bivalent strain to either replicate in vivo or colonize the mucosal surfaces.
The following examples illustrate the practical problems encountered in the construction of bivalent vaccine strains. Formal et al. (Infect. Immun. 34 (1981), 746-750) have introduced the 120 Mdal virulence plasmid of S. sonnei into S. typhi Ty21a via conjugation. The resulting hybrid strain, termed 5076-1C, expressed the O-PS antigen of S. sonnei encoded by the plasmid on the surface of Ty21a as a capsular-like material unbound to S. typhi LPS core (Seid et al., J. Biol. Chem. 259 (1984), 9028-9034). Immunization of volunteers with this strain resulted in a vigorous anti-S. sonnei LPS antibody response. However, in challenge studies, various lots of this vaccine were unable to consistently afford significant protection against S. sonnei disease (Herrington et al., Vaccine 8 (1990), 353-357; Black et al., J. Infect. Dis. 155 (1987), 1260-1265; Van De Verg et al., Infect. Immun. 58 (1990), 2002-2004). The precise reason for this variable protection has not been identified. Possible explanations include, 1) the presence of the S. sonnei antigen on the surface of the Ty21a strain interfered with its ability to effectively colonize, 2) the virulence plasmid was shown to be genetically unstable within Ty21a giving rise to spontaneous deletions which interfered with the expression of the S. sonnei O-PS and other virulence-associated antigens, 3) expression of the S. sonnei plasmid in Ty21a could have led to a deleterious effect manifested only in vivo such as reduced survival, multiplication or colonization.
A bivalent vaccine strain was constructed by introducing the genes encoding for V. cholerae O-PS biosynthesis into Ty21a yielding strain EX645. This strain induced a modest anti-V. cholerae LPS immune response when fed to volunteers even though the heterologous O-PS was coupled to the LPS core (Forrest et al., J. Infect. Dis. 159 (1989), 145-146). Only a modest level of protection was afforded against experimental cholera following immunization with EX645. Subsequent studies showed that the longer S. typhi O-PS probably masked the somewhat shorter V. cholerae O-PS units accounting for the poor immune response. A derivative of EX645, termed EX880, was developed by inactivating genes involved in the expression of the S. typhi O-PS. EX880 was found to induce a far more vigorous anti-V. cholerae LPS antibody response compared to EX645 (Attridge et al., Infect. Immun. 59 (1991), 2279-2284). The anti-S. typhi LPS response was minimal.
The rfb/rfc and the rfaR1 loci of S. sonnei were introduced into CVD103-HgR by the use of compatible plasmids (Viret et al., Mol. Microbiol. 7 (1993), 239-252). This allowed for the efficient expression of the S. sonnei O-PS coupled to LPS core. However, when these same genetic loci were introduced into the chromosome of CVD103-HgR (strains CH3 and CH9), little if any S. sonnei O-PS was covalently coupled to LPS core (Viret and Favre, Biologicals 22 (1994), 361-372). Instead, the material was expressed on the surface of CVD103-HgR as a capsular-like material.
The above observations suggest the following, 1) heterologous O-PS can be efficiently coupled to homologous or heterologous LPS core only if the synthesis of homologous O-PS is suppressed, 2) under appropriate conditions it may be possible to covalently couple heterologous O-PS to the unique core of V. cholerae thereby obviating the need for introducing genes coding for a heterologous core molecule, and 3) the co-expression of two distinct O-PS molecules by the same carrier strain resulting in a bivalent vaccine may not be feasible. Thus, the efficient simultaneous expression of two complete LPS molecules each presenting different O-PS moieties may be beyond the capacity of a single host strain. Possible reasons include interference with the expression of the respective genes at the transcriptional level, competition for limiting components involved in the biosynthesis of the outer membrane structure, such as molecules involved in the transposition of the O-PS molecule to the outer surface of the cell, or competition between the O-PS molecules for transfer or binding to available sites on the LPS core molecule.
In an attempt to circumvent these problems previously spontaneous, undefined mutants of V. cholerae CVD103-HgR which are deficient in the synthesis of O-PS were isolated. Such strains were capable of supporting the covalent attachment of S. sonnei O-PS encoded by the chromosomally integrated rfb/rfc loci to an LPS core. However, the undefined nature of the mutation(s), present in such strains render them unacceptable for human use.
Thus, the technical problem underlying the present invention is to provide live attenuated vaccine carrier strains, which are useful for the expression and delivery of heterologous O-antigen (O-PS) from gram-negative bacteria in such a way that the heterologous O-PS can induce an immune response and which are safe and acceptable for administration as a vaccine.
The solution to said technical problem is achieved by providing the embodiments characterized in the claims. It has been surprisingly found that a defined genetic modification can be introduced in a live attenuated vaccine strain, which does not interfere with the functions of the carrier strain required in order to make said strain suitable as carrier for a heterologous antigen, and which leads to a deficiency of said strain in the synthesis of homologous O-PS, thereby allowing to efficiently express a desired heterologous O-PS in such a manner that the heterologous O-PS is covalently coupled to the LPS core and can induce an immune response.
The embodiments of the present invention inter alia allow for the construction of monovalent vaccine strains with the following characteristics, 1) use of a live oral attenuated vaccine strain, preferably V. cholerae CVD103-HgR, suitable for human use as a carrier for heterologous antigens, 2) modification of said carrier strain so as to render it deficient in the synthesis of homologous O-PS by introduction of precise mutations, e.g. within the rfb gene which are non-lethal, halt the synthesis of homologous Inaba O-PS and allow for the expression and covalent coupling of heterologous O-PS to the LPS core, 3) containing genes necessary for the production of heterologous, polymerized LPS molecules derived from other enteric pathogens and expressing them, wherein stable expression is achieved by integration of the cloned heterologous genes at a site which does not adversely affect the phenotype of the carrier strain, specifically, those traits which would allow it to induce a protective immune response following oral administration, 4) expression of the heterologous O-PS genes in such a manner that the encoded O-PS is covalently coupled to either the LPS core of the carrier strain or a heterologous LPS core produced by the carrier strain following the introduction of the appropriate rfa locus, 5) the LPS molecule bearing the heterologous O-PS moiety is expressed on the surface of the carrier strain, preferably integrated into the outer membrane protein, and 6) the genotype/phenotype of the carrier strain which renders it suitable for human use is maintained.
In order to develop such vaccine strains various genetic modifications were introduced in the genes for expression of the O-PS of the carrier strain in order to eliminate synthesis of the O-PS.
Surprisingly, the deletion of the entire Inaba rfb locus (about 20 kb) had a lethal effect upon CVD103-HgR and its S. sonnei rfb/rfc-bearing derivatives (strains CH3 and CH9). Therefore, it was assumed that there must be genes encoding for essential functions within or adjacent to the rfb locus and that strains deficient in such functions would be unable to multiply, presumably due to their inability to synthesize a functioning outer membrane structure. It was therefore sought to introduce specific deletions, for example, within three distinct regions of the rfb locus. The goal was to try to introduce non-lethal deletions into the rfb locus which would, in addition to halting expression of the homologous Inaba O-PS moiety, support the covalent coupling of the heterologous S. sonnei O-PS to the Inaba LPS core. The first such construct was a rfbEGHI mutant. The rfbE locus encodes the perosamine synthetase whereas the rfbG, H, and I loci are involved in the transport of the Inaba O-PS through the outer membrane (Manning et al. p.77-94. In Vibrio cholerae and Cholera: molecular to global perspectives (1994). Wachsmuth K., Blake, P. A., and Olsvik xcfx86. (eds.). Washington, D.C.:American Society for Microbiology). The rfbEGHI deletion was found to be lethal in CVD103-HgR. In a second strain, deletion of the rfbN locus (which is involved in the synthesis of the perosamine substituent 3-deoxy-L-glycero-tetronic acid), unexpectedly resulted in only weak production of the heterologous S. sonnei O-PS which was unbound to the Inaba core. Therefore specific gene functions within the Inaba rfb locus are useful for both the expression of the heterologous S. sonnei rfb genes and its covalent coupling to the V. cholerae LPS core in as of yet unidentified manner. Next the rfbA and rfbB loci were inactivated by deleting a 1.2 kb fragment overlapping the junction between the two loci. These loci are involved in the synthesis of the perosamine component of the Inaba O-PS. Specifically, the RfbA protein is associated with enzymes having phospho-mannose isomerase or mannose-1-phosphate guanyl transferase activity while the rfbB loci encodes a putative phospho-manno mutase. The introduction of the rfbA/rfbB mutation into CVD103-HgR containing the S. sonnei rfb/rfc loci allowed for the expression and covalent coupling of the S. sonnei O-PS to the Inaba LPS core giving rise to full length hybrid LPS molecules. Recombinant strains expressing the Inaba rfbA/rfbB deletion together with the S. sonnei rfb/rfc loci with or without the R1 core were found to be genotypically and phenotypically stable upon passage in vitro. Furthermore, these strains possessed all the characteristics of the CVD103-HgR strain which render it suitable for human use, including, 1) lack of cholera toxin activity, 2) production of non-toxic B subunit of cholera toxin, 3) expression of toxin co-regulated pili, and 4) the ability to grow in the presence of elevated levels of Mercury ions.
Accordingly, the present invention relates to live attenuated vaccine strain against gram-negative enteric pathogens characterized by the following properties:
(a) deficiency to express homologous O-PS due to a, defined genetic modification, and
(b) capability to efficiently express heterologous O-PS in such a manner that said heterologous O-PS is covalently coupled to the LPS core.
As used herein, the term xe2x80x9cdefined genetic modificationxe2x80x9d encompasses any modification(s) which has (have) been introduced by recombinant DNA techniques and which is (are), in contrast to modifications introduced by random mutagenesis or due to spontaneous mutations, defined with respect to its nature and location. Said modifications can be deletions, additions, substitutions or rearrangements of nucleotides, but should preferably not give raise to the occurrence of revertants. Suitable genetic modifications in accordance with the present invention can be introduced by the person skilled in the art following the teaching given in the Examples below. Such modifications should not interfere with the functions of the carrier strain required in order to make said strain suitable as carrier for a heterologous O-PS, but should sufficiently eliminate the expression of homologous O-PS. For example, said modifications affecting the biosynthesis of the homologous O-PS should not adversely affect the expression of genes which are essential for the synthesis of complete LPS comprised of heterologous O-PS, e.g. the genes involved in the synthesis of lipid A, the LPS core, the synthesis and transport of O-PS to the outer cell surface and anchoring the LPS molecules into the outer membrane.
Due to said modifications said strains synthesize LPS molecules which only consist of the homologous lipid A and homologous and/or, in a specific embodiment which is described below, a heterologous LSP core. Preferably, said modifications are deletions. As used herein, the term xe2x80x9cdeficiency to express homologous O-PSxe2x80x9d means that the expression of the homologous O-PS is entirely eliminated or at least reduced such that the efficient expression of the desired heterologous O-PS, and its covalent binding to the LPS core of the carrier strain or, alternatively, to a heterologous LPS core is made possible.
As used herein, the term xe2x80x9ccapability to efficiently express heterologous O-PSxe2x80x9d means the capability to express said O-PS in such a way, that the amounts of heterologous O-PS produced are sufficient to elicit an immune response.
In a preferred embodiment, the vaccine strain carries a defined genetic modification within the genes involved in O-PS biosynthesis contained in the rfb, rfc, and/or rfp loci or any combination thereof.
In a particularly preferred embodiment, the vaccine strain carries a defined genetic modification within the rfbA-, rfbB-, rfbD and/or rfbE-gene or any combination thereof, preferably within the rfbA- and/or rfbB-gene.
Most preferred is a vaccine strain, wherein said genetic modification is a deletion corresponding to the deletion shown for pSSVI255-20 (DSM depository number DSM13426) in FIG. 1. This deletion is located at the beginning of the rfbinaba operon and concerns the elimination of a 1.2 kbp HindIII fragment. It inactivates the rfbA- and rfbB-genes which are involved in the biosynthesis of the perosamine O-antigen subunit.
Suitable vaccine strains can be selected by the person skilled in the art, depending on the desired purpose. Such strains are, for example, CH19 (DSM depository number DSM13420), CH21 (DSM depository number DSM13421), CH22 (DSM depository number DSM13422), CH24 (DSM depository number DSM13423), CH25 (DSM depository number DSM13424) or CH30 (DSM depository number DSM13425), described below.
In a preferred embodiment, said vaccine strain is an E. coli strain, a strain of the genus Shigella, S. typhi, O1 or O139 V. cholerae, Heliobacter pylori or Campylobacter jejuni. Preferred S. typhi strains are S. typhi Ty21a, S. typhi CVD908, or S. typhi CVD908 containing additional attenuating mutations. Examples of additional attenuating mutations are mutations in the viaB or htpR genes encoding transcriptional signals such as the RpoS sigma factor or in genes involved in virulence traits such as the resistance to environmental stress or the capacity to adapt to new growth conditions or in genes involved in the synthesis of aromatic acids.
Preferred V. cholerae strains are V. cholerae CVD103-HgR, V. cholerae CVD103-HgR, CVD110, CVD111, CVD112, Bengal-15 or Peru-14.
Preferred Shigella strains are S. dysenteriae, S. sonnei, S. boydii, or S. flexneri serotype Y.
The above vaccine strains can be used for the efficient expression of heterologous O-PS. For this purpose a heterologous gene or a set of heterologous genes coding for O-PS are inserted into the vaccine strain by methods known to the person skilled in the art, for example by methods described in the Examples, below.
Accordingly, the present invention relates to vaccine strains further characterized by the presence of a heterologous gene or a set of heterologous genes coding for O-PS.
The insertion of said gene(s) encoding a heterologous O-PS should be carried out in such a manner that (i) said gene(s) are stably expressed and allow for the synthesis of complete full-length, smooth LPS essentially indistinguishable from the parent strain, and (ii) an intact hybrid LSP is formed composed of the lipid A of the vaccine strain coupled to the homologous core region. Thus, when inserting said gene(s) the person skilled in the art should
i) use a bacterial carrier strain devoid of the genes coding for the homologous O-PS,
ii) use a plasmid, for example pMAK700oriT, composed of all the genes coding for the heterologous O-PS, flanked by homologous genetic regions corresponding to the locus where the said heterologous O-PS genes are to be inserted, and
iii) then proceed as described in Example 3, below.
In a preferred embodiment of the vaccine strains, the heterologous gene(s) is (are) present either on a plasmid vector or stably integrated into the chromosome of said strain at a defined integration site which is to be non-essential for inducing a protective immune response by the carrier strain.
The set of heterologous genes should be cloned in a deletion vector composed of a thermosensitive replicon, for example, pMAK700oriT and a homologous genetic region corresponding to the gene where the insertion is to take place. The heterologous genes will be cloned in the middle of the homologous region. For the integration of the heterologous genes this plasmid should be introduced into a suitable carrier strain and thereafter handled like in Examples 5, 6, 7 and 8, below.
Suitable sites for integration of the heterologous gene(s) into the chromosome of the vaccine strain are genes which in no way will effect properties of the strain necessary for its immunogenecity and safety.
In a preferred embodiment, said heterologous gene or set of heterologous genes are integrated into either the hlyA, hlyB, rfbA, and/or rfbA/rfbB loci of V. cholerae. 
A further particular preferred embodiment relates to a S. typhi strain, wherein said heterologous gene or set of heterologous genes are integrated into either the H2S production gene, ilv, viab, htpR genes encoding transcriptional signals such as the RpoS sigma factor, genes involved in virulence traits such as the resistance to environmental stress or the capacity to adapt to new growth conditions, or any gene involved in the synthesis of aromatic acids. Genes involved in the resistance to environmental stress or the capacity to adapt to new growth conditions are genes of the OmpR-EnrZ system, PhoP-PhoQ system and cya-crp transcription regulation system. Genes involved in the synthesis of aromatic acids are, for example, aroA, aroC and aroD.
Alternatively, the above vaccine strains contain the rfa, rfe, rfp, and/or any additional gene(s) necessary for the synthesis of complete smooth heterologous LPS which are integrated in tandem into a single chromosomal site or independently integrated into individual sites.
Additional genes necessary for the synthesis of complete smooth heterologous LPS are for example, rfc and rff. Integration of the above genes in such a way that they are correctly and in a coordinate manner expressed can be carried out by the person skilled in the art according to well known methods or, for example, described in Hamilton et al., J. Bacteriology 171 (1989), 4617-4622.
Such vaccine strains allow expression of heterologous O-PS which is covalently coupled to a heterologous LPS core region, which, preferably, exhibits a degree of polymerization essentially indistinguishable from that of native LPS produced by the enteric pathogen. Such vaccine strains can, if desired, modified in such a way that they are deficient in the synthesis of homologous LPS core.
In a preferred embodiment, the heterologous rfa genes encode the Ra, R1, R2, R3, R4, K-12 or B LPS core, preferably the R1 core.
The invention also relates to a live vaccine comprising the above vaccine strain and optionally a pharmaceutically acceptable carrier and/or a buffer for neutralizing gastric acidity and/or a system for delivering said vaccine in a viable state to the intestinal tract.
Said vaccine comprises an immunoprotective and non-toxic amount of said vaccine strain. Suitable amounts can be determined by the person skilled in the art and are typically 107 to 109 bacteria.
Pharmaceutically acceptable carriers, suitable neutralizing buffers, and suitable delivering systems can be selected by the person skilled in the art.
In a preferred embodiment said live vaccine is used for immunization against gram-negative enteric pathogens.
The mode of administration of the vaccines of the present invention may be any suitable route which delivers an immunoprotective amount of the vaccine to the subject. However, the vaccine is preferably administered orally or intranasally.
The invention also relates to the use of the above vaccine strains for the preparation of a live vaccine for immunization against gram-negative enteric pathogens. For such use the vaccine strains are combined with the carriers, buffers and/or delivery systems described above.